Quantum-Inspired Chemical Rule for Discovering Topological Materials

Area of Science:

• Materials Science
• Quantum Physics
• Artificial Intelligence

Background:

• Topological materials possess unique electronic structures crucial for quantum phenomena and advanced technologies.
• Discovering new topological materials is hindered by computationally expensive calculations and slow experimental synthesis.
• Existing machine learning methods, like the topogivity rule, offer data-driven prescreening but lack quantum insights.

Purpose of the Study:

• To develop a novel quantum-inspired machine learning approach for efficient topological material discovery.
• To overcome the limitations of classical heuristics by incorporating quantum-native features.
• To enhance the predictive power and physical interpretability in topological material classification.

Main Methods:

• Developed a hybrid quantum-classical neural network (HQCNN) integrating compositional descriptors with quantum probability amplitudes.
• Formulated a quantum-inspired rule that naturally captures interelement correlations.
• Validated the physical consistency and interpretability using an equivalent complex-valued neural network (CVNN).
• Employed high-throughput screening combined with density functional theory (DFT) calculations.

Main Results:

• The HQCNN successfully maps compositional data to quantum probability amplitudes, revealing interelement correlations.
• The quantum-inspired rule provides efficient and generalizable topological classification.
• High-throughput screening identified five previously unreported topological compounds.
• The approach demonstrates enhanced predictive power compared to classical heuristics.

Conclusions:

• The developed quantum-inspired heuristic offers a powerful and efficient tool for discovering novel topological materials.
• This method bridges chemical intuition with quantum mechanical principles for materials prediction.
• The findings pave the way for accelerated exploration of topological quantum matter and its applications.